Key maturity benchmarks for priority practices that can improve IT productivity.
Choose a practice to learn more:
Developer journeys and engineering practices
Not exhaustive
Modern engineering practices
Fully reliant on requirement specification and quality assurance in product delivery
Limited use of development practices such as feature flagging or A/B tests to reduce risk and better validate product changes
Wide use of development practices (eg, feature flagging,
A/B tests, blue-green and canary deployments) for product changes
Higher profitability
Preferred delivery methodology
80% or more of delivery in IT follows waterfall delivery methodology
A mix of waterfall and iterative delivery methodologies, with some areas preferring one over the other
80% or more of delivery in IT follows iterative delivery methodology
Higher profitability
Application patterns
No application archetypes are defined; greenfield development is started from scratch
Teams define their own templates that cover basic code structure and functionality
Automatically generated boilerplates use enterprise application patterns with
logging/monitoring
Software delivery process
Mostly manual; different teams have their own accommodations for routine tasks
Around 50% of applications onboarded to CI/CD pipelines, enabling deployment automation; limited integration of automated guardrails, stage gates, or controls
Automated code deployment from development to production (with embedded testing/cyber controls); more than 80% of applications onboarded to subversion software and target pipelines
1 of 3
Product and platform operating model
Not exhaustive
Cross-functional delivery teams
Higher profitability
Most delivery teams are functional with only technology capabilities (eg, developers, quality assurance)
50–60% of the delivery teams possess business, technology, and operations capabilities enabling them to deliver and maintain products
80% or more of the delivery teams include full BizDevSecOps capabilities and are able to deliver and maintain their product end to end
Vendor dependency
Higher profitability
More than 80% of applications are delivered by vendors/contract development
30–50% of applications are delivered by vendors/contract development
Less than 10% of applications are delivered by vendors/contract development
Number of reporting levels
Higher profitability
Typically 5 or more levels of hierarchy between a developer and the CEO
Typically 4 levels of hierarchy between a developer and the CEO
Typically 3 or fewer levels of hierarchy between a developer and the CEO
AI use case delivery teams
Higher profitability
Opportunistic AI exploration in select domains (eg, workforce management, risk)
Centralized AI pipeline or funnel in place but not fully linked with the business strategy; 20–50 AI use cases in production
AI-enabled use cases embedded in business processes/operational improvements, with defined performance metrics; first proofs of concept on gen AI; more than 100 use cases in production
2 of 3
Application simplification and decoupling
Not exhaustive
Decoupling and service independence
80% or more of teams depend on other teams during releases; there are known monolith systems that are clear bottlenecks in scalability
Around 50% of teams depend on other teams during releases; there are not more than 1–2 monolith systems that are bottlenecks in scalability
95% or more of teams can make releases independently from other teams; each service scaled independently subject to actual loads
Capability/application duplication
Higher profitability
More than 30% of applications in the application portfolio have overlaps in functionality
10–20% of applications in the application portfolio have overlaps in functionality
There are no applications with overlapping functionalities; all applications are aligned with business capabilities
Omnichannel capabilities
Higher profitability
Products/services for different channels are implemented in separate systems, with no shared code base; business logic is specific to a channel
Notable effort toward delivering products/services across different channels, with limited code reusability
Business logic and process orchestration are largely reused between different channels
Public cloud workloads
Higher profitability
Less than 10% of workloads on public cloud
30–40% of workloads on hybrid cloud; unified control plain across on-premises and public cloud
60% or more of workloads on hybrid or multicloud
3 of 3
Less developed
Advancing
Mature
McKinsey & Company